When is AI going to regenerate the lost data on recordings from cd or other digital source?

[smoking_man]

A sound field is like the surface of the ocean but in 3 dimensions. There are many waveforms in that surface just as there are in acoustic compression waves.

 

[smoking_man]

If you cannot capture or model all those waves then you cannot reproduce the original sound recording.

 

[dfuller]

If you cannot capture or model all those wave then you cannot reproduce the original sound recording.

But you can. Quite easily, in fact.

 

[Joe Smith]

HA, I can see the marketing pitch now. Every damn album from the 60s, 70s and 80s remastered yet again, no matter how obscure, and sounding "like you've never heard it before..."

Remember the lesson from WarGames: "The only way to win...is not to play."

 

[Cbdb2]

a musical performance does not just create one wave, like cd. It produces a multitude of waves which are simplified down to one waveform when recorded.

No, a musical performance creates two waves (signals), one at each ear and so does a CD and headphones.

 

[tineared]

I would be very interested in seeing what AI could do for Stevie Ray Vaughn's Live At El Mocambo concert video

 

[Ropeburn]

The clicks and pops perhaps?

I used to record and save vinyl in 24/96 at a point. Dirty Techno vinyl worn by DJ use, no less. My obsessed arse even did manual declicking using headphones and looking at the waveform, then repairing the clicks by redrawing the waveform by hand. Then normalise to full volume. Took about an hour for one track. Autodeclickers are flawed machine algorithms for amateurs!

 

[DVDdoug]

You are correct that the spatial reproduction is imperfect. That's NOT a limitation of the 16-bit, 44.1KHz format. it's more related to the limitations of stereo, and the fact that you are probably in a small room and you can hear the small room acoustics. A pair of speakers might fool you in a concert hall. (Experiments/demonstrations with speakers in a concert hall have been done.)

But surround sound can be very good!!!

I have a shelf-full of concert DVDs (and a few Blu-Rays) with surround sound but they are mostly rock so there is no attempt to create the sound of a concert hall or even the actual sound that was heard in the actual venue. (And live rock music is usually mono from the PA system.)

With regular stereo music, I use a "hall" or "theater" setting on my AVR for some delayed reverb in the rear to simulate the sound of a larger space. It's not AI, and I'm NOT listening "as intended". Most music is "artificially" created in the studio and mixed/mastered on a pair of stereo speakers/monitors. It IS possible to get virtually perfect "studio sound" at home! ...My setup isn't THAT good and my living room is not acoustically treated. But it still sounds good to me. And with the up-mixed surround effect I'm not trying to get the original studio sound.

AI can create near perfect video

It's still a "video". It's not a 3-D hologram that fools you into thinking you're seeing a live orchestra in front of you! A movie (even in 3-D) doesn't look like a stage play.

 

[antcollinet]

I used to record and save vinyl in 24/96 at a point. Dirty Techno vinyl worn by DJ use, no less. My obsessed arse even did manual declicking using headphones and looking at the waveform, then repairing the clicks by redrawing the waveform by hand. Then normalise to full volume. Took about an hour for one track. Autodeclickers are flawed machine algorithms for amateurs!

It's madness. Madness, I tell you.

"One Step Beyond"

 

[Ropeburn]

It's madness. Madness, I tell you.

"One Step Beyond"

It is. Today I got wiser, dig in my record collection, and then listen to proper YouTube uploads of the same tracks. Sounds so much better, and I'm saving money for new cartridges and especially turntables. My Vestax PDX2000 have some of the most stable and reliably tracking tonearms ever made... the downside is them being straight and very short with big angle error and audible resonance... especially on clicks. Lol!

 

[waofner]

AI can create near perfect video from its data set including voice and background sound, I think it would be able to manage to create near perfect audio from nothing which does not require video.

seems you are begging to be lied to.

 

[Audionaut]

The 16bit 44.1khz standard looses musical information from the original performance, this cannot be recreated by an ordinary DAC. Will AI eventually fill in the gaps?

This will not be a reconstruction of the original, but rather an AI hallucination.
But probably better than the 80% that AI-generated music currently accounts for in new music titles.

 

[Vandemann]

I want an AI agent that analyzes all of the sound information in a recording, including room reverberation and spatial locality, and reproduces it in pristine form. Think Disklavier pianos but using digital samples instead of real instruments.

 

[mhardy6647]

i won't be satisfied until the ai can recreate the oboist's funky breath and the second violist's uncomfortably tight shoes, stuff like that

That's what Dolby®™© Atmos©®™ is for... isn't it?

 

[Ropeburn]

I want an AI agent that analyzes all of the sound information in a recording, including room reverberation and spatial locality, and reproduces it in pristine form. Think Disklavier pianos but using digital samples instead of real instruments.

That's the problem right there.

What is "pristine"? Who determines that?

As a, let's say, seasoned amateur and dilettante musician and producer, I make the decisions. My brain and taste say what's good or not. What's original and what's fake. What's a dry enough signal? I can decide, so can every other amateur, and experienced professionals doubly so.

The AI, what reference and experience does it have, beyond a very superficial faux-understanding of the whole?

 

[antcollinet]

I want an AI agent that analyzes all of the sound information in a recording, including room reverberation and spatial locality, and reproduces it in pristine form. Think Disklavier pianos but using digital samples instead of real instruments.

All of the information in a recording is already reproduced in pristine form.


It's only the information that didn't make it into the recording that can't be reproduced. Not even by AI.

 

[smoking_man]

There are only 88 notes on a piano, it will not take very long for AI to reproduce any piece of music. There are only 8 notes in an Octave. How many permutations of music can you produce with 8 notes? AI will be able to create a large finite number of recordings with those 8 notes. It's Rocket Science or probably not.

 

[Nopiano]

There are only 88 notes on a piano, it will not take very long for AI to reproduce any piece of music. There are only 8 notes in an Octave. How many permutations of music can you produce with 8 notes? AI will be able to create a large finite number of recordings with those 8 notes. It's Rocket Science or probably not.

I think it’s mathematics. But very big numbers for any normal duration of ‘song’.

Multiply by the different instruments in the orchestra with different pitch, overtones and timbre, and you understand why music is more art than science.

 

[antcollinet]

There are only 88 notes on a piano,

And yet there are an infinite number of ways those 88 notes can be combined over a 3 minute song - when you include timing, intensity, sustain, combination with other instruments and voices and etc. etc. etc.

It is why over the centuries since the piano and similar instruments were invented it is still possible to create hundreds?/thousands?/10ks? songs per day that are unique.

 

[smoking_man]

well, music is mainly repitition. You have the chorus and the verse. A plain example why music is limited in its current version. I would have thought that statistics can show the number of musical permutations that you can have, if I remember correctly it is 8x7x6x5x4x3x2x1=40,320 musical combinations.